Natural experiments

Natural experiment occurs when some exogenous event (e.g., government policy) changes conditions for some group. They are used to uncover “causal relationships,” and often require advanced statistical analysis. 

Angrist (1990) provides an example. This paper aims to estimate how military service affects lifetime earnings. It is often hard to estimate such effect directly because people without good working opportunities are usually more likely to serve in the military, so the lower-income is not “caused” by military service. 

The paper uses a government policy during the Vietnam War – eligibility for military service is “randomly” determined by the draft lottery using the dates of birth. In this case, it could guarantee that when conducting statistical analysis, all the other characteristics of the veterans and the non-veterans are comparable, so we can be sure that the difference in income is “caused” by the military service. The analysis shows that the earnings of veterans were, on average, about 15 percent less than the earnings of non-veterans.

Angrist, J. D. (1990). Lifetime Earnings and the Vietnam Era Draft Lottery : Evidence from Social Security Administrative Records. American Economic Association, 80(3), 313–336.